Oracle is making a calculated play to loosen Nvidia’s grip on the AI chip market.
The cloud giant will deploy 50,000 AMD AI accelerators in its infrastructure starting in 2026. The move positions Oracle as a growing rival in the increasingly competitive AI cloud space, where companies are racing to secure high-performance chips to power generative AI workloads.
“We feel like customers are going to take up AMD very, very well — especially in the inferencing space,” said Karan Batta, senior vice president of Oracle Cloud Infrastructure, according to CNBC.
This week’s announcements unveil two bold moves: Oracle is boosting its hardware arsenal with AMD and launching a new data platform that embraces open standards.
Oracle ups the ante in chip wars
Oracle said Tuesday it will begin deploying 50,000 of AMD’s MI450 AI accelerators in Q3 2026, with plans to scale beyond that as production ramps up.
The initiative strengthens Oracle’s position in the fast-growing AI infrastructure market, providing customers with a new option in a space currently dominated by Nvidia’s GPUs. AMD’s stock rose more than 3% following the announcement, reflecting growing confidence in its ability to compete at scale.
Oracle’s investment is part of a broader alignment with AMD that echoes the industry’s push for hardware diversity. The company’s upcoming Oracle Cloud Infrastructure (OCI) expansions will offer customers both AMD and Nvidia GPU instances, enabling more flexibility for AI model training and inference.
It’s also closely tied to Oracle’s deepening relationship with OpenAI, under a reported $300 billion, five-year deal that will see OpenAI using Oracle’s cloud infrastructure to run AI workloads.
Still, questions remain about the economics of the deal. Oracle and AMD have not disclosed cost structures or procurement timelines, leaving analysts to speculate on how quickly the new capacity will come online and what margins Oracle can achieve.
More must-read AI coverage
- SS&C Intralinks DealCentre AI vs. Datasite: Which platform is built for the future of dealmaking?
- SS&C Intralinks FundCentre AI vs. Juniper Square: Which platform better supports modern private markets fund managers?
- Why Data, Not Models, Determines AI Success
- The Rise of the AI-Native Factory: How Physical AI Is Transforming Manufacturing
Oracle launches open lakehouse, leans into interoperability
Hardware wasn’t Oracle’s only headline this week.
The company also introduced the Autonomous AI Lakehouse, a next-generation data and analytics platform designed to unify data management across clouds while supporting Apache Iceberg. This open table format allows easier data interoperability.
The new platform builds on Oracle’s Autonomous Database, combining AI-driven data preparation, SQL query acceleration, and built-in governance. According to InfoWorld, the service integrates with external catalogs like Databricks Unity, AWS Glue, and Snowflake Polaris—an open, multi-vendor ecosystem that reflects Oracle’s shift toward vendor-neutral interoperability.
Under the hood, Oracle enables GoldenGate data replication into Iceberg tables and exposes them through the Autonomous Database catalog. The system supports query engines like Trino, Presto, and Spark, effectively eliminating friction in hybrid analytics environments.
With its dual push into AMD-powered infrastructure and open data architecture, Oracle is reshaping how enterprises think about cloud-scale AI. The company’s challenge now lies in execution — proving it can balance performance, interoperability, and cost in a market where customers have more choice than ever.
Also worth reading: Oracle’s stock soared nearly 40% after announcing multibillion-dollar AI cloud contracts. Dive deeper into how CEO Safra Catz and Larry Ellison are positioning Oracle’s inference stack as the next trillion-dollar battleground.